Meta AI 眼镜：下一代计算平台

如果技术发展继续沿着其轨迹前进，确保我们拥有积极未来的最佳方式是什么？

What is the best way to ensure that we have a positive future if the technology development continues on its trajectory?

2004 年，哈佛大学柯克兰楼的一个年轻人有了一个想法。痴迷于个人计算与社交网络交汇点的马克·扎克伯格将他当时的个人编程项目“The Facebook”上线。这一小举动彻底改变了社交媒体，并永远改变了全球文化。在随后的几年里，扎克伯格将 Facebook 的受众和影响力扩大到超出任何人最疯狂的想象。他随后利用这一成功巩固权力，整合了包括 Instagram、WhatsApp 和 Oculus 在内的一系列其他平台，最终将母公司更名为 Meta。在这一飞速崛起过程中，扎克伯格和 Meta 践行了公司“快速行动，打破陈规”的创始精神。然而，胜利总是超过失误，22 年后，这个曾经的颠覆者已成为全球性机构。2026 年 6 月 23 日，在纽约市，我们与扎克伯格进行了一场 Idea Generation 的特别现场录制，讨论这一切，以及他最新市场领导者产品——Meta AI 眼镜的发布。而这一切都始于一个想法。

In 2004, a young man in Harvard University's Kirkland House had an idea. Obsessed with the intersection of personal computing and social networking, Mark Zuckerberg pushed his personal programming project, then known as The Facebook, live. And in that small act, revolutionized social media and changed global culture forever. In the years that followed, Zuckerberg would scale Facebook's audience and influence beyond his or anyone's wildest imagination. He would then leverage that success into a consolidation of power, rolling up a litany of other platforms, including Instagram, WhatsApp, and Oculus, eventually renaming the parent company Meta. Throughout this meteoric rise, Zuckerberg and Meta would live up to the company's founding ethos, moving fast and breaking things. However, the wins would always outpace the missteps, and 22 years later, the once upstart iconoclast is now a global institution. On June 23rd, 2026, in New York City, we sat down with Zuckerberg for a special live taping of Idea Generation to discuss all of this, as well as the launch of his latest market leader, Meta's AI glasses. And it all started with one idea.

大家好。欢迎各位。感谢大家参加 Idea Generation 的现场录制。我想不言而喻，地球上没有哪位创意企业家比坐在我身边的这位先生——马克·扎克伯格——更负责塑造我们共同分享的现实。

Hello. Welcome, everyone. Thank you for joining us for this live taping of Idea Generation. I think it goes without saying that there is not a creative entrepreneur on the planet who has been more responsible for crafting the reality that we all share than the gentleman that sits next to me, Mark Zuckerberg.

感谢你加入我们。

Thank you for joining us.

是的，这对我们来说是大日子，所以感谢你来做这个。时机正好。我们对今天的发布非常兴奋。新的 Meta 眼镜系列是我们与合作伙伴 EssilorLuxottica 从头设计的第一款眼镜。这建立在我们与雷朋、奥克利等标志性品牌的合作基础上，基本上将所有功能带到了新系列中，我们希望它能被更多人使用，同时包括我们构建的最先进 AI 模型直接集成到眼镜中。所以，我们非常高兴能这样做。

Yeah, well, this is a big day for us, so thank you for doing this. This is great timing. You know, we're super excited for the announcements today. The new line of Meta glasses, it's the first line of glasses that we've designed with our partner EssilorLuxottica from the ground up. This builds on the work that we've been doing with Ray-Ban and Oakley and iconic brands and it basically brings all of that functionality to a new line that we hope will be accessible to a lot more people, but brings all of the functionality including the most advanced AI models that we've built right into the glasses. So, we're very happy to be doing this.

眼镜这种形态的哪些方面吸引你，让你认为它可能是手机的下一代版本？

What is it about the form factor of glasses that draws you to think that will be the next version of the phone maybe?

是的，总体上有几件事让我们对眼镜成为下一个计算平台感到乐观。一是随着时间的推移，计算变得越来越自然，越来越融入我们的生活，而眼镜让你能够与技术互动，同时保持与周围人的联系，我认为这是非常根本的一点，显然手机做不到。当你与手机互动时，你是在与这个小矩形互动。它把你从周围世界中拉走。你希望技术能够有效地完全融入你周围的世界。所以，这是很重要的一点。另一点是，眼镜是让 AI 助手看到你所见、听到你所听、全天与你交谈，并最终在你的视野中显示信息、叠加在周围世界上的理想形态。其他形态只能实现部分功能。如果你有耳机，也许它能听到你所听，也许能与你交谈。但我认为视觉部分将非常重要。显然，眼镜是这方面的理想形态。还有一点是，眼镜是人们已有的东西，对吧？世界上已有近 20 亿人佩戴光学眼镜。还有数十亿人戴太阳镜。所以，我认为我们现在正处于这样一个时刻：就像每个人都用翻盖手机，然后智能手机出现了。在 5 年内，所有翻盖手机都将变成智能手机，这一点变得非常清楚。这就是我今天对眼镜的感觉。5 年后，人们佩戴的任何眼镜为什么不能拥有这种全天帮助人们的功能呢？这不太合理。所以，我认为这是一个非常自然的设计，人们已经习惯了。

Yeah, so I guess there's a few things overall that make us optimistic that glasses are the next computing platform. One is that you have this trajectory over time where computing gets to be more natural and more integrated in our lives and glasses just give you the opportunity to interact with technology but remain present with the people around you, which I think is a really fundamentally important thing that obviously phones don't do. When you interact with your phone, you're kind of interacting with this small rectangle. It pulls you away from the world around you. You want to have technology be able to effectively fully integrate with the world around you. So, that's a big one. Another piece is that glasses are the ideal form factor for you to be able to give an AI assistant that works for you the ability to see what you see, hear what you hear, talk to you throughout the day and then eventually also can display information in your field of view and be able to overlay things on the world around you. So, that's something that you can get parts of that with other form factors. You might be able to, if you have earbuds, maybe it can hear what you hear, and maybe it can talk to you. But, I think the visual part is going to be very important. And obviously, glasses I think are the ideal form factor for that. And then one more thing is just that glasses are a thing that people have, right? I mean, there's almost 2 billion people in the world wear optical glasses already. And billions more wear sunglasses. So, I kind of think we're at this moment now where it's kind of like, everyone had flip phones, and then you got smartphones came out. And it just kind of became pretty clear that in 5 years or whatever, all of the flip phones were going to be smartphones. And that's basically how I feel about glasses today. It's like, what is the reason why 5 years from now any of the glasses that people wear are not going to be able to have this kind of functionality to help people out throughout their day. It just doesn't quite make sense. So, I think it's a very natural design that people are used to.

你是什么时候有这种顿悟的？一旦你抓住这个想法，你需要开始设置哪些多米诺骨牌？

How long ago did you have this revelation? And sort of, once you seize upon this idea, what are the dominoes that you have to start setting up?

嗯，我想我们从 2014 年就开始做这个了。对吧？我们最早的工作是从虚拟现实和增强现实开始的。想法是，虚拟现实允许你在进入人们每天佩戴的形态之前，以稍大的形态进行一些世界增强。但我们一直在做的眼镜，到现在已经超过 10 年了。我认为与 EssilorLuxottica 的合作，我们从最初的雷朋开始做了好几代。然后每年我们都增加更多功能。现在，你可以打电话、听音乐。它拥有帮助视力不佳的人的功能。如果有人听力不好，它可以放大耳朵听到的声音。就像这样，一代一代地增加各种功能。AI 也在不断改进。我认为我们还会再研究几十年，但未来 5 年将是关键时期。

Well, I mean, we've been working on this I think since 2014. Right? I think the earliest part of our work on, we started working on virtual reality and augmented reality. And the notion is that, okay, virtual reality allows you to do some of the augmentation of the world in a somewhat larger form factor before you get into the one that people are going to wear every day. But, some of the glasses that we've been working on it's more than 10 years at this point. I think that the partnership with EssilorLuxottica, we've done multiple generations starting with the original Ray-Bans that we did. And then it just each year we've added more functionality. It's gotten to, you know, it adds more functionalities. So, now, you know, you can do phone calls, you can listen to music. It has all this functionality to help people who might not be able to see. If someone can't hear well, it can amplify the hearing in their ears. Like, all the different functionality you just kind of add generation over generation. The AI keeps on getting better. Now, I think we're going to be working on this for decades more, but I think the next 5 years are going to be a really key period for this.

你使用这些眼镜的时间可能比地球上任何人都长。它们如何改变了你的日常生活？

You have had access to these probably longer than almost anyone on the planet. How have they changed your daily routine?

我注意到的主要是，有很多事情你可以在手机上做。但如果你戴上眼镜，效果更好。比如，能够接电话，耳朵是开放的，没有被遮挡。所以，就像我走在街上，你能听到周围发生了什么。我觉得这真的很棒。

The main thing that I notice is just there are a lot of things that you could do on your phone. But if you have glasses on, it's just better. All right, so it being able to talk to like take phone calls and have your ear open and not occluded. So, it's like, all right, I'm just walking down the street and you can like hear what's going on. I think that's really nice.

麦克风藏在鼻托里，拾音清晰度极高，就算你站在风洞里，对方听起来也完全清晰。我甚至在水上摩托上开过商务电话——不常干，但确实发生过，对方根本听不出来。这很棒，对吧？你也不想让对方知道你在骑水上摩托。所以我觉得这类细节很好，但真正让眼镜成为主流平台的方式是，其他计算平台并不会消失。我们开始用手机时，电脑并没有消失，我们仍然有电脑。但不知道你们有没有这种感觉，我经常坐在桌前，电脑就在旁边，却还是伸手拿手机去发消息，因为手机已经成了主要平台。我认为眼镜也会如此。随着眼镜能做越来越多的事，我们会进入一个世界：你仍然带着手机，但它更多时候只是待在口袋里，最终可能更多时候被留在家里。但我不认为未来 10 年我们会停止使用手机，只是每一代眼镜都会越来越成为我们使用的主要设备，越来越融入我们的生活。

The fact that the microphone is in the nose pad, so the clarity of the audio that it picks up is so high that you could literally be in a wind tunnel and it would sound completely clear to the person on the other side. I mean, I've taken business calls on like a jet ski. Not often, but it has happened. And the other person could not tell that I was on it. It's great, right? It's like, that's good. You don't necessarily want to tell the other person that you're on a jet ski. So, I think that that kind of stuff is nice, but the main way that I think that this is going to work is that, you know, it's not like all the other computing platforms go away, right? When we started using phones, our computers didn't go away. We still have computers, but I don't know how many of you feel this way, but I'm often sitting at my desk, my computer is right there, but I still just reach for my phone to go send someone a message or something, right? Because it's just the phone has kind of become the primary platform. And I think similarly, glasses are going to be that. So, I think as you can do more and more with the glasses, we'll just kind of be in a world where you'll have your phone with you, but it'll just kind of stay in the pocket more of the time. And then maybe eventually you'll leave it at home more of the time. But I don't think that, you know, in the next 10 years we're going to stop having phones or something like that. I just think that glasses are going to each generation become more and more the primary thing that we use and, you know, just be more integrated into our lives.

从 2014 年开始做这个，我知道 Meta 在 AI 浪潮中起步很早，在 2010 年代负责了很多早期的机器学习工作，远早于在座各位与 LLM 交互。过去 6 到 12 个月发生了革命性的进步。

Having started this in 2014, I know that Meta is very early or it has been very early on the AI wave and, you know, responsible for a lot of early machine learning in the 2010s, long before anyone in this crowd was interacting with an LLM. There have been revolutionary steps that have taken place in the last 6 months, 12 months.

完全同意。

Totally.

这如何改变了你对眼镜内部功能以及人们使用方式的看法？

How has that changed sort of how you think about what is inside the glasses and how people will use them?

一个巧妙之处在于，如果你把硬件做对了，硬件本身迭代很慢。我们规划这一代眼镜大概用了两年，现在已经在规划 2028 年的产品线了。但在这期间，AI 会不断升级，越来越好。如果你今天买了这副眼镜，随着 AI 越来越强大、能做更多事情，它会变得越来越有用。过去几年最大的进步，简单来说是多模态理解。你不仅能提问，还能让它看到你周围的情况，并跟你讨论，这非常有用。语言翻译现在几乎已经解决了，眼镜内置了实时翻译。你去另一个地方，它能直接跟你说话并翻译另一种语言。但我最兴奋的是智能体式 AI 的进步。我的意思是，从一问一答的模式，转向你给它一个任务，它自己去思考，过一会儿再回来；或者像实时会话那样，比如我在跟女儿一起烤东西，我完全不在行，只是为了陪三岁的女儿玩，它可以直接说“别那样做，这样做”，我甚至不用问它。它一直在看着，知道情况。比如“别把面粉放那儿”。这真的很棒。随着你调整系统，你可以从半小时的常开 AI 延长到一小时、两小时，最终全天候。那将是一种完全不同的 AI 使用模式。当然你可以控制它何时开启，但很快我们就能让 AI 看到你所见、听到你所听，并全天候帮你总结、协助。当你把 AI 接入这些数据时，你会发现我们一天中错过了多少信号——我们因为专注而忽略了其他重要的事情，或者有人在对话中说了什么。拥有这种记忆或能力，以各种方式帮助你，将变得极其重要。

So, one of the neat things is that if you get the hardware package right, hardware is kind of slow, right? So, we've been planning like this generation of glasses for whatever, 2 years, whatever the cycle is for that, and you know, we're planning the 2028 line of glasses now, right? And all that. Then you get all these releases in between that are basically just the AI getting better and better. So, if you buy these glasses today, they're just going to keep on getting more and more useful for you as the AI becomes more powerful and able to do more things. So, the biggest advances in the last couple of years, I'd say there's the simple one on multimodal understanding, right? So, just being able to not just ask it a question, but actually have it be able to see what's going on around you and be able to kind of talk to you about that is quite helpful. And obviously language translation at this point is almost solved, right? So, you have the real-time translation built into it. So, you travel to another place, it can basically speak to you and then translate another language. But one of the things that I'm really excited about is the advance in the agentic AI. So, what I guess I mean by that is moving away from this use pattern where you ask it a question and it's one question, one answer, right? It's what I think this is going to evolve more towards is either you give it a task and it goes away and thinks and then comes back a bit later, or you just have these like live sessions where, you know, it's like I'm cooking, I'm like baking something with my daughter and I'm terrible at this, and I'm just trying to do it to hang out with my 3-year-old who's into baking, and like it can basically just it's like no, really don't do that. Like do this instead. And it's like, okay, that's good. I didn't even have to ask it. It just it's watching, right? It knows. Like do not put the flour there. So, that I think is really good. Right? And basically as you tune the system, you can go from, you know, half an hour of the always-on AI to an hour to 2 hours to eventually that's like just all day. Right? And then that's a very different mode of using the AI where now, obviously you'll have control over how it is, when you want it to be on and not, but you can see that like pretty quickly I think we're going to get to the zone where you effectively have the ability for your AI to see what you see, hear what you hear, summarize that, help you out with all of these things throughout the day, and it's just kind of amazing when you plug this stuff into an AI, how much of the signal that we have throughout our day that we just kind of miss or we, you know, we tune stuff out because we're focused on something, but there's something else happening that was important, or someone said something in a conversation, and you know, so it's kind of the ability to have that kind of memory or be able to help you in all these ways is going to be incredibly important.

你如何让它变得正常化？因为它确实能让你更投入，但也在某种程度上侵入你的生活，尤其是当它主动参与你的一切时……

How do you sort of normalize, because this does allow you to be present, but it is somewhat intrusive in your life and particularly if it is helping you and actively engaging in everything that you...

你当然要确保。我们的设计理念是：眼镜首先必须是一副好眼镜。意思是技术是次要的，时尚和佩戴体验才是首要的。作为一家科技公司，我们意识到这一点很有意思，因为大多数科技公司想的是“我们把所有技术塞进去，形状看起来像眼镜，为什么人们不戴？”但现实是，你会戴你喜欢的眼镜。你戴眼镜很多小时，但只会在其中一小部分时间里使用技术。这非常重要。为了解锁这种使用，眼镜必须让你引以为傲，并且舒适，这样你才会长时间佩戴。所以首先必须是一副好眼镜。这也意味着要“不碍事”。很多科技公司的愿景是“我们这里能帮你，那里能帮你”，然后不停地轰炸你，但人们不想要那样，我也不想要。

You want to make sure. I think so, our mantra for designing the glasses is that they have to be great glasses first. I guess what we mean is that the technology is secondary, the fashion and the wearing it is actually primary, which is kind of an interesting thing for us to get there as a technology company because I think most of the technology companies think about it is like, all right, we're trying to cram all this technology into it. It's like this shape looks kind of like the glasses that they put out. Like why wouldn't people wear this? But I think the reality is that you're going to wear the glasses you like. Whatever glasses you have, you're going to wear for many hours and you're going to use the technology for some small fraction of it. And it's actually really important. So in order to unlock that, it needs to be that the glasses are something that you're proud to wear and they need to be comfortable so that way you'll wear them for long periods of time. So they have to be great glasses first. Part of that also means getting out of the way. Right? So I think a lot of technology companies kind of have the vision of oh well we can help you here and we can help you here and we can help you here and like okay let's just kind of bombard you with stuff and it's like people don't want that. I don't want that.

对吧？我觉得你其实希望它在打扰你或打断你的时候非常克制。所以我们就是这么想的：首先是出色的眼镜，当你需要时提供技术，功能越来越多，但我们真的不希望它变得侵入性。我去年秋天看你接受了一个采访，大概是雷朋最新款发布的时候。你当时在谈论你对 AI 的整体愿景，试图阐述这样一个想法：会有定制化的模型，人们可以策划自己使用的一套 AI。一年后，这个领域有了很多创新，我很好奇，你现在还是这么看吗？

Right? It's like I think you actually want it to be very sparing in terms of when it will ping you or interrupt you on something. So that's kind of the way that we think about it is great glasses first, technology available when you want it with more and more capabilities, but we really don't want it to be intrusive. I was watching an interview that you gave I guess last fall around the release of the latest Ray-Bans. And you were just sort of talking about your vision for AI in general and how you were trying to articulate I guess the idea that there would be sort of bespoke models then that people would sort of curate their own sort of suite of AI that they use. And I'm curious a year later because there has been quite a bit of innovation in the field. Is that still how you see it all playing out?

是的。我们在 AI 方面的愿景，我们称之为个人超级智能。这个想法我认为与其他所有 AI 实验室的共识观点非常不同。我认为大多数其他从事 AI 的公司认为会有一个大 AI 供所有人使用，或者它是在为世界做事。也许这只是我们公司从社交媒体建设历史中继承的遗产。一切都是关于构建工具，把它们交到人们手中，相信人们应该有使用工具和表达他们想要的东西的自主权。从根本上说，我认为为了拥有一个强大 AI 的美好未来，你需要很多制衡。好吧，我不想生活在一个只有一个大 AI 的未来。我认为那是一个糟糕的未来。无论 AI 有多好，我认为那都不好。我认为最终变得好的方式是引导技术发展去赋能每一个个体。现实是，个体有不同的愿望。这些愿望有时重叠，有时相互竞争。这种竞争很重要，因为它是我们社会和市场的制衡方式。我认为在一个每个人都拥有个人超级智能的世界里，它拥有你想要的上下文，因为你拥有各种设备可以给它关于你生活的上下文。它可以帮你实现目标，24/7 为你工作。全世界尽可能多的人拥有它，对我来说，这是我们获得真正美好未来的最佳机会。对吧？在那个世界里，我认为人们被赋能，但同时你也有一个基本的制衡，权力平衡。所以我认为这是我们的哲学中非常基本的东西。我们开发的一切，从 AI 软件、模型和技术，到硬件、眼镜以及我们制造的不同产品，让人们以不同方式使用它，都是为这一哲学服务的。

Yeah. So our vision for what we're trying to build on the AI side we call personal superintelligence. And the idea I think it's very different from what the kind of consensus view of all of the other AI labs is. I think most other companies that are working on AI think that there's going to be like one big AI that everyone uses or that it's kind of like doing stuff for the world. I mean maybe this is just a legacy that we have from the history of the company building social media. It's all about building tools, putting them in people's hands, believing that people should have agency to use the tools and express the things that they want to. In that there's fundamentally this view I don't know, I think that in order to have a good future with a strong AI, you want a lot of checks and balances. All right, so I don't want to live in a future where there's like one big AI. I think that's a bad future. No matter how good the AI is, I think that's not good. I think the way that this ends up being good is you actually just direct the technology development towards empowering every individual. And the reality is that individuals, people have different aspirations. And the aspirations sometimes overlap with each other and sometimes they compete with each other. And that competition is important because it is sort of our society and markets' way of putting checks and balances on the way that things work. And I think in a world where kind of everyone has a personal superintelligence that has the context that you want it to have because you have whatever devices you have that can give it the context that you want it to have about your life. And it can help you achieve your goals and work on your behalf 24/7. And as many people around the world as possible have that, to me that's the best shot that we have of getting to a future that is like really good. Right? It's like in that world I think people are empowered, but also you kind of have a fundamental check on this and there's a balance of power and how that works. So I think that is just a very fundamental thing for our philosophy. And everything that we're developing from the kind of AI software and the models and the technology to the idea of the hardware and that the glasses and different things that we'll make to make it so people can use this in different ways. It's all in the service of that philosophy.

你提到了制衡及其重要性。我不知道你是否熟悉去年秋天出版的一本书，《如果有人建造它，每个人都会死》。我有点好奇。你认为这种关于 AI 影响的末日担忧有任何可信度吗？

You mentioned the sort of checks and balances and the importance of that. I don't know if you're familiar with the book that came out last fall, 'If Anyone Builds It, Everyone Dies'. I'm a bit curious. Do you think that there's any credence to that kind of doomer concern around the implications of AI.

我想我不太熟悉那本书，但我认为 AI 安全的概念显然非常重要，对吧？所以我比那要乐观一点。我不认为如果有人建造它，每个人都会死，但我确实认为任何强大的新技术都会赋予人们做好事的能力，也会赋予他们做坏事的能力。你给人们的杠杆越强大，他们可能带来的好处和伤害就越大。所以，认真对待这一点，努力减轻伤害，同时尽可能为人们带来积极影响，是极其重要的。但这就回到了我所说的哲学：如果技术发展继续沿着它的轨迹，确保我们拥有积极未来的最佳方式是什么？我认为有些人会说，‘好吧，你要做的是构建一个单一的 AI，并使其尽可能对齐。’我只是不知道这是否足够，对吧？因为现实是，如果你有一个超级强大的东西，我不知道。我的意思是，历史上有很多人或政权获得了绝对权力，通常结局都不好，对吧？无论那个东西多么开明。我认为人类在将开明构建到事物中的能力是有限的，无论我们多么努力。所以我有点认为人类和文明有一个自然机制，那就是制衡。所以，我创造积极未来的版本不是有一个中心化的东西，而是你需要广泛分布它。这在某种程度上是一个有争议的概念，因为有些人会说，‘哦，如果你给个体一个具有很多能力的模型或 AI，他们不会做坏事吗？’但我不知道。我认为这始终是自由与赋能个体之间以及如何建设社会之间的历史张力。我认为在每一步，当我们尝试给予人们更多自由和自主权，而不是说‘哦，这对一个人来说太多了’时，我们往往会受益。我们希望人们拥有这样做的个体自主权。现在，我认为那是通往更美好未来的道路。所以，我们这样做。我对未来相当乐观。但，我不知道。

I guess I'm not as familiar with that book, but I think the concept of AI safety is obviously a very important one, right? So, I guess I'm a little more optimistic than that. I don't think if anyone builds it everyone will die, but I do think that anything that is a powerful new technology gives people the ability to do good things and it can give people the ability to do bad things. And the more powerful the lever that you're giving people, the more good and the more harm that they can potentially do. So, obviously taking that very seriously and working to mitigate the harm while making it as positive as possible for people is incredibly important. But that's where you get to this philosophy that I was talking about which is what is the best way to ensure that we have a positive future if the technology development continues on its trajectory? I think some people would say, 'Okay, what you want to do is build a singular AI and make it like as aligned as possible.' And I just I don't know that that's going to be enough, right? I think because the reality is this, if you have one thing that is super powerful, I don't know. I mean there have been a lot of people or regimes that get absolute power over time and that usually just doesn't end well, right? No matter how enlightened the thing is. And I think humans are limited in our ability to build enlightenment into a thing no matter how hard we try. So, I kind of think that there is a natural mechanism that humanity and civilization has which are checks and balances. So, my version of this in terms of creating a positive future is not that you have one central thing, but I think it's like you really need it to be widely distributed. And it's in some ways it's somewhat of a controversial concept because I think some people would say, 'Oh, well if you give a model or an AI that has a lot of capabilities to individuals, won't they do bad things?' But I don't know. I think that this has always been a historical tension between liberty and empowering individuals and in how you build society. And I think at every step along the way we tend to benefit when we try to go for giving people more freedom and more agency and not saying, 'Oh, this is too much for a person to handle.' We want people to have the individual agency to do that. Now, I think that that's the path to a better future. So, I think we do that. I'm pretty optimistic about the future. But, I don't know.

不同的人试图建造不同的东西。所以，我希望我们的愿景能够实现。

The different people are trying to build different things. So, I hope our vision is the one that gets built.

我很好奇，因为，你知道，你在我们所知的社交媒体的创建中显然起了关键作用，并且显然处理了第二和第三阶后果。有些令人惊讶且积极，有些则必须被缓解。

I'm curious because, you know, you were obviously instrumental in the creation of social media as we know it and obviously dealt with second and third-order consequences. Some surprising and positive, some that had to be mitigated.

关于 AI，你认为有哪些即将到来的事情是人们没有预料到的，会让社会更有活力，或者带来潜在危险？

With AI, what are the things on the horizon that you think people are not expecting that will sort of enliven society or present potential dangers?

嗯，我想目前大多数人关注的是 AI 拥有了巨大的编程能力，这很有帮助，因为即使不是工程师的人，也有很多编程任务希望 AI 能完成。比如，我在我的 MMA 拳馆里装了摄像头，然后说，‘好吧，你能看着我在回合之间给我反馈吗？’这其实就是一个编程问题，对吧？因为它现在需要处理视频管道，处理所有进来的图像帧，传输进来，判断我什么时候在回合之间，然后给我发送信息。所以，显然，编程能力越强，它就越能应对——你知道，你会面临更多的网络安全威胁。我认为这显然是现在的一个大话题。我认为即将到来的一件事是，我们会开始看到更多的科学进步。所以，我在 Meta 之外的慈善工作，我和 Priscilla 创建了 BioHub，我们的目标是帮助加快科学研究的步伐，以帮助科学家在本世纪治愈和预防所有疾病。我们刚开始时觉得这非常雄心勃勃。好吧，在本世纪治愈和预防所有疾病。现在看来可能太保守了。我不知道是十年还是二十年，但我想会比本世纪末快得多。我认为这在很大程度上是因为构建这些 AI 模型的进展，它们能够有效模拟生物学，从而帮助药物设计，帮助理解生物学的基本部分以及如何让这一切达到正确的平衡。但是，你知道，能够创造这些的东西，我担心的是，会不会有坏人用它来制造伤害人的东西？或者制造生物武器之类的。我认为这仍然是一个悬而未决的问题，但显然需要仔细思考。我认为这是即将到来的一件事，我们需要小心。所以，我认为未来一两年内，这两件事有一条解锁巨大价值的路径，也有清晰的路线图来缓解一些担忧，但也有相当大的担忧需要小心。然后，一旦超越那个阶段，我认为就会进入更广泛的社会问题，比如我们刚才谈到的权力平衡，我认为这些将是非常根本的。你知道，人们谈论诸如工作被取代的焦虑。公司总是试图更高效地做事，对吧？我相信你也是，对吧？如果你有一个任务，你不想一直重复做同样的事。你会想办法更高效地完成。同时，人们也在提高自己的生产力。这有一个自然的平衡，对吧？所以，如果公司自动化东西的速度比个人提高生产力的速度快，那么你可能会认为工作数量会减少，但人们认为这是不可避免的。我实际上不这么认为。我认为如果你专注于赋能人们，让他们更高效，而且这个速度比公司自动化的速度快，那么理论上未来的工作应该更多，而不是更少。所以，这就是为什么你需要这种平衡，对吧？如果只有少数公司专注于自动化所有知识工作，那么那可能不是一个美好的未来，对吧？但如果有一个平衡，一些公司专注于让公司更高效地工作，而另一些则专注于这种个人超级智能的愿景，即赋能个人，让每一步都更高效，那么我认为结果可能会相当不错。所以，我不知道。我认为这些都是需要仔细思考和小心对待的事情，但我认为有一条非常积极的前进道路。

Yeah, so I guess the ones that most people are focused on right now is you're having this huge ability for the AI to be able to code, which is helpful because even for people who are not engineers, there's just a lot of coding that you want it to be able to do. So, it's like I put cameras up in my MMA gym and I'm like, 'All right, can you just watch me and give me feedback in between rounds?' And it's like, 'All right, that's actually a coding problem.' Right? Because now it has to deal with the video pipeline of processing all the image frames that come in and piping that in and figuring out when I'm in between rounds and then sending me the thing. So, obviously the better that stuff gets at coding, the better it gets at being able to, you know, you get more cybersecurity threats. I think that's obviously one of the big topics now. I think one of the things on the horizon that we're going to start seeing is more scientific advances. So, my philanthropic work outside of Meta, Priscilla and I created this thing BioHub where our goal is to help advance the pace of scientific research in order to help scientists cure and prevent all diseases this century. And when we started that we thought that was really ambitious. Okay, like cure and prevent all diseases this century. Now it's probably too conservative. I don't know if it's going to be a decade or two decades or whatever, but I would guess it'll be much sooner than the end of the century when you get there. And I think a lot of that is because of the progress of building these AI models that can effectively simulate biology and therefore help with drug design and help with understanding of fundamental parts of biology and how to get the right equilibrium for how all that works. But, you know, the things that can create that, I think there's the concern that well, could someone who's bad use that to make things that hurt people, right? Or create like a bio weapon or something. And I think that's still an open question, but I think it's clearly a thing that really needs to be thought through carefully. And I think it's one of the things on the horizon that we need to be careful about. So, I think that those are probably the two in the next year or two that I think there's a path to unlock a huge amount of value and also I think a clear roadmap to mitigate some of the concerns, but also pretty big concerns that I think you want to be careful about. And then, once you get beyond that, I think you start getting into these broader societal questions around like the balance of power and stuff that we talked about a minute ago that I think are just going to be really fundamental. You know, people talk about stuff like, obviously there's a lot of anxiety about job displacement. Companies are always trying to do things more efficiently, right? It's like I'm sure you do, right? If you have a task, you don't want to just keep on doing the same task. You try to find a way to do it more efficiently. And at the same time, people kind of improve their own productivity. Well, there's a natural balance of this stuff, right? So, if companies get very good at automating stuff faster than individuals get better at productivity, then you'd kind of assume that the number of jobs might decrease, but I think that people assume that that's an inevitability. I don't actually think it is. I think if you focus on empowering people and making people more productive, and that happens at a faster rate than companies get better at automating things, then in theory there should be more jobs in the future, not less. So, I think that's why you need this balance, right? If you just have a few companies and they're focused on automating all knowledge work, then that's probably not a great future, right? But if you have a balance where some companies are focused on making it so that companies can work more efficiently, but others are focused on more of this personal superintelligence vision where you're empowering individuals and making people more productive at each step along the way, then I think it's probably going to be pretty good. So, I don't know. I think that this is just all stuff that you want to think through and be careful about, but I think that there is a path forward that can be very positive on this.

嗯，你提到了网络安全问题，显然每个人都在谈论 Anthropic 不得不撤回 Mythos 的事。我很好奇，因为你显然曾作为顾问与联邦政府合作过。你认为我们应该如何推进这些前沿模型？

Well, you mentioned the cybersecurity thing, which obviously everyone's talking about with Anthropic having to roll back Mythos. And I'm curious because you've obviously worked with the federal government as an advisor. How do you think we should move forward with these sort of frontier models?

你知道，不同的担忧有不同的答案。但就拿我们刚才讨论的工作问题来说，我的理念是，更多的答案来自于把权力交到更多人手中，而不是限制它。有一个光谱：有一群人会说，‘这非常危险。我们必须是唯一拥有它的人。’另一群人会说，‘不，这很强大，也有一些问题，但答案是让它广泛可用，人们会有效地相互制衡。’我认为不同的问题在这一点上会有不同的平衡。但我认为我们公司的本能和哲学更倾向于赋能个人，认为如果你把权力交到更多人手中，你会得到更好的状态。我绝对认为在工作问题上这是对的。在网络安全上可能也是对的，对吧？我的意思是，想想开源软件，它往往更安全。这有点反直觉，对吧？早期人们认为，‘好吧，开源，人们能看到代码，这意味着他们能更容易发现漏洞。这肯定更不安全，对吧？’但反直觉的是，如果更多人能看到代码，然后人们发现漏洞，你修复漏洞，然后你可以广泛地推出补丁，最终你会得到更坚固的软件。所以，我认为对于所有这些，你想要的就是快速加固它。

You know, different concerns have different answers for them. But, you know, if you take the jobs one that we were just talking about, my philosophy on this tends to be that more of the answer will come from putting the power into more people's hands rather than holding it back. There's kind of a spectrum from there're going to be a set of people who are going to argue, 'This is really dangerous. We must be the only ones who have this.' And then there's a set of people who will say, 'No, this is powerful and there's some issues, but the answer is to make it so that this can be widely available and people will effectively check each other.' I think different issues will have different balances of where you want to be on that. But I think our instincts and philosophy as a company are more on the side of empowering individuals and thinking that more of the answers you get to a better state if you basically put the power in more people's hands. I definitely think that that's true on jobs. I think that's probably true on cybersecurity, right? I mean, if you think about stuff like open-source software, it tends to be way more secure. It's somewhat counterintuitive, right? I mean, people thought early on, 'Okay, open-source, like people can see the code, so that means that they can find the vulnerabilities more. This has got to be less safe, right?' But kind of counterintuitively, if more people can see the code and then people find the vulnerabilities, then you fix the vulnerabilities, and then you can roll out the open-source like you can roll out the patches for that widely, and then you end up with much more hardened software. So, I kind of think for all of this what you want is just harden it very quickly.

我希望我们能达到这样一个阶段：借助 AI，你可以编写可验证的安全代码，AI 基本上能让人们写出代码——无论是刚才提到的家庭自动化，还是他们在自己生活中做的任何事情——并且他们基本上可以让 AI 帮助他们编写可验证的安全代码，从而基本上消除大部分这些问题。

I hope that we will get to a point with AI where you can write verifiably secure code that the AI can basically make it so that people can write code, whether it's a home automation thing or whatever they are doing in their own lives, and they can basically have the AI help them write code that is verifiably secure, so that you kind of eliminate most of those issues.

你对 AI 下了相当大的赌注——我记得是一年前对 Scale AI 投资了 140 亿美元，并聘请了 Alexander Wang。你对 Meta 在 AI 竞赛中的现状满意吗？一年后你计划达到什么位置？

You made a fairly large bet on AI with your, I think it was a $14 billion investment in Scale AI and hiring Alexander Wang a year ago. Are you happy with where Meta is in the AI race and where do you plan to be a year from now?

我的意思是，我从来不会对现状满意。所以，我觉得……这对我来说效果如何？不，我是说，你得看看前排的那些家伙。就像‘今天不错，你推出了好东西。明天回去工作。’现在，不，我是说，我们取得了很大进展。我们有一系列 Llama 模型，但不幸的是 Llama 4 模型没有达到我们需要的轨迹。所以我们做了件大事，重启实验室并重建它。如果你一年前问我，我们新建的实验室到现在还不到一年。我记得是 6 月 30 日我们启动了 Meta 超级智能实验室。如果你告诉我我们在模型进展上能达到今天的位置，我会非常高兴。但因为一路上我已经习惯了这些好消息，我现在觉得我们应该做得更好。但这更多是我的性格。不过行业正在取得良好进展。我对此很满意。我对眼镜也很满意。我认为这不仅仅是实验室里构建模型的问题。问题是，如何让 AI 真正在人们的生活中发挥作用。我们已经有了几种形态。手机或电脑上的 AI 助手应用，当你使用时很酷，显然非常有用。有超过十亿人在使用。但我只是觉得，我们希望达到一个更环境化、融入生活的状态。我认为我们学到的一点是，目前很多 AI 产品还不够个性化。它们开始从之前的对话上下文中理解一些东西。但现实是，关于你的大部分信息并不存在于你与 AI 的对话中。我认为，如果你想实现个人超级智能的愿景，让 AI 能在一天中环境化地帮助你，像眼镜这样的设备最终会变得非常关键。这就是我们长期致力于此的原因之一。我认为这是我们最引以为豪的事情之一，这是一个非常新颖的智力贡献，因为我认为没有多少其他公司真正专注于将前沿 AI 与真正好的设计融合，做成你想佩戴的东西。

I mean, I'm never happy with where we are. So, I think that that's... How is it working for me? No, it's... I mean, you got to look at the guys in the front row. It's like, 'It's a pretty good day. You launched something good. Back to work tomorrow.' Now, no, I mean, look, we're making a lot of progress. We've had these run of the Llama models, and unfortunately the Llama 4 model was not on the trajectory we needed to be. So we did this pretty big thing to reboot the lab and kind of build it up. And if you'd asked me a year ago, the new lab that we built is less than a year old at this point. I think it was June 30th that we started the Meta Super Intelligence labs. If you told me that we'd be where we are today in terms of model progress, I would have been very happy with that. But because I have acclimated to the good news along the way, I now think that we should be doing even better. But that's kind of more my personality. But the industry is making good progress. I'm quite happy with it. I'm quite happy with the glasses. I think it's not just about building a model in a lab. I think the question is what are the ways that this can actually be something that is useful in people's lives. And we have a few form factors for this already. You have the AI assistant app on your phone or on your computer, and that's cool when you're doing that. Obviously very useful. You have like a billion people or more doing that. But I just think that you want to get it to a point where it's more ambient and woven into your life. And I think one of the things that we've learned is that a lot of the AI products that are out there today are not very personalized yet. They're starting to understand some stuff from the context of previous conversations. But the reality is that most of the stuff about you doesn't exist in conversations that you've had with your AI. I think if you want to get to this vision around personal superintelligence and something that can ambiently help you throughout your day, I think devices like glasses are going to end up being really critical to that. So that's one of the reasons why we've been working on this for a long time. And I think this is one of the things that I'm most proud of that we're doing, it's a very novel intellectual contribution, in that I don't think that many other companies are really focused on fusing frontier AI with really good design that are something that you just want to wear.

你目前在 AI 辅助眼镜市场占有约 80% 的份额。最近一周左右，一些竞争对手发布了公告，进入了这个领域。作为一个相当有竞争力的人，你怎么看他们？

You presently have about 80% of the market share in the AI-assisted glasses. Some competitors have stepped into the ring in the last week or so with announcements. What are your thoughts about them as a fairly competitive person?

是的，我是说，看，现在还为时过早。所以，我认为很多其他公司都会这么做，因为 Meta 和 EssilorLuxottica 已经证明了这是可行的。

Yeah, I mean, look, it's early. So, I think a lot of other companies are going to do this because Meta and EssilorLuxottica have shown that this works.

是的。

Yes.

对吧？所以，我认为我们尝试过很多不成功的事情，对吧？而在这个领域，我认为它显然有效，现在当你有了这样的东西，行业自然会关注它。我们在这方面有相当大的技术领先优势。所以我对未来几年即将推出的东西感到非常兴奋，但还有很多东西需要发明，对吧？所以，我们这里讨论的一部分是能看、能听、能跟你说话的眼镜。显然还有显示方面的视觉部分。所以，我们去年开始推出带有小显示屏的 Meta Ray-Ban 显示器。我们已经展示了全宽视场显示器的原型版本。

Right? So, I think we've tried a lot of things that didn't work, right? And basically in this one, I think clearly works, and now it is natural and normal that when you have something like that, the industry will focus on that. And we have quite a technological lead on this. So I feel really excited about the stuff that's coming up over the coming years, but there's a lot to invent, right? So, part of what we're talking about here, the glasses that can see and hear and talk to you. There's obviously the visual side of it with the displays. So, we started rolling out the Meta Ray-Ban displays last year with the small display. We've shown the prototype version of the full wide field of view display.

太不可思议了。

It's incredible.

我的意思是，它很神奇。它比我们想要的要重一些……我们试图达到……

I mean, it's magical. It's a little heavier than we'd like it to be for... we try to get to like...

它相当令人印象深刻。

It's pretty impressive.

谢谢。我的意思是，那个原型我们花了大约 10 年时间。它很酷。我们想把它降到大约 50 克，也许 60 克，甚至可能先到 70 克，但我们认为全天舒适度的真正目标可能接近 50 克。而且，人们想要各种不同的款式。人们喜欢不同的东西。有些人喜欢非常薄的眼镜，有些人喜欢更厚实的眼镜。厚实的眼镜有助于放入更多技术，对吧？但这也有限制，对吧？人们会想要一个光谱上的不同东西——有些人会想要非常少的技术，非常轻。然后有些人会想要完全沉浸式的宽视场全息图。你可以坐在咖啡店里，面前有三个巨大的显示器；或者你在飞机上，有一个巨大的宽视场电视显示屏，它物理上不存在，但就在那里，那会很棒。但是的，我认为人们会在功能上想要光谱上的不同点。还有不同的价格点。你放入的技术越多，它就会越厚，我们认为我们可以让它时尚，但需要一些时间才能把微型化做得很好。而且放入的技术越多，价格就越贵。所以人们会根据时尚、功能和价格点来做决定。但这里有很多事情要做，我认为我们采取了一种相当极简主义的方法……我想我们只是认为眼镜不是手机。人们不想要和手机一样的砖块。

Thank you. I mean, that one we worked on for like 10 years, that prototype. It's cool. We want to get it down to about 50 grams, maybe 60 grams, or maybe even 70 grams to start, but the real target for what we think is all-day comfort is probably closer to 50 grams. And you also mean people want a variety of different styles. Like people like different things. So some people like very thin glasses, some people like chunkier glasses. Chunky glasses are helpful for putting a lot of technology into, right? But there are limits on that too, right? And there's going to be this spectrum of things that people are going to want from some people I think will want kind of very minimal technology, very light. And then some people here are going to want the full immersive wide field of view holograms. You can sit down at a coffee shop and have like three huge monitors up in front of you, or you're on a plane and you just have a huge wide field of view TV display that doesn't physically exist, but it's there, and that's going to be awesome. But yeah, I think that there are going to be different points on the spectrum that people want for what functionality they want to have. And also different price points. So the more technology you put into it, the thicker it will be, which we think we can make stylish, but that will take some time to get the miniaturization to be very good. And also the more technology you put in, the more expensive it is. So people will make decisions based on fashion, functionality, and the price point that they want. But there is a lot to do here, and I think we're taking a pretty maximalist approach of trying to... I guess we just think that glasses are not a phone. People do not want the same kind of brick as a phone.

你提到这些尝试有时会失败。你知道，在你的职业生涯中，你下了很多大赌注，从收购 Instagram 和 WhatsApp 到元宇宙，媒体都在事后评头论足。我很好奇，你认为可接受的成败比例是多少？

You mentioned these working and trying things sometimes that don't work. And you know, you've made a number of large bets during the course of your career, all of which have been backseat driven by the media from, you know, acquiring Instagram and WhatsApp to the metaverse stuff. I'm curious, what do you think is the sort of acceptable ratio of working to not work?

我不知道这是不是个比例。

I don't know if it's a ratio.

没人能百发百中。

Nobody bats a thousand.

嗯，是的，我知道，当然不是。我认为问题在于你赢的时候赢了多少。就像沃尔特·艾萨克森写爱因斯坦传记时，我觉得最有趣的一点是他那些错误的理论。当你回顾他的一生，人们往往不记得那些。对吧？我觉得我们都会犯很多错。重要的是我们做出的智力贡献或文化贡献，推动事物向前发展。所以，我不知道。我觉得有很多类比。你可以选一个格斗类比。就像，你被打倒很多次。真正重要的是你尝试了多少次，对吧？站起来继续前进。我认为我们是一家非常有韧性的公司，不太害怕失败。这让我们能够大胆尝试，也能长期忍受被嘲笑，因为某种程度上，如果你在做真正新的事情，人们必须在一段时间内觉得它荒谬，否则别人早就做成了。它已经在奥弗顿窗口里了。所以，你本质上是在做人们觉得奇怪或有点疯狂的事。你显然会尽量脚踏实地。但我有点觉得，如果你不推动，人们不觉得你在做的事有点狂野，那么你可能只是没有足够努力，没有用力挥棒。这个回答混了 18 个比喻，但你知道。

Well, yeah, I know, definitely not. I think the question is how much you win when you win. Like Walter Isaacson wrote this biography of Einstein. And like one of the things that I thought was the most interesting is like all the theories he had that were wrong. And it's like when you look back on his life, people don't tend to remember those. Right? It's like I think it's like we're all going to get a lot of stuff wrong. I think what matters are the intellectual contributions or, you know, cultural contributions that we make that help push stuff forward. So, I don't know. I think that there's like a lot of analogies. You can, you know, pick a fighting analogy or something. It's like, all right, you get knocked down a bunch. Like what really matters is how many times do you keep trying, right? Get up and keep going. I think we have a pretty resilient company that is not very afraid of failure. And that allows us to take some big swings and also tolerate being ridiculed for long periods of time because to some degree, if you're trying to do something that's really new, people have to think it's ridiculous for some period of time because if they didn't, then someone else would have already done it. It would already be in the Overton window. So, you're fundamentally trying to do things that people think are like weird or a little crazy. And you know, you obviously try to do it in a way that is as grounded as possible. But I kind of think if you're not pushing it and people don't think what you're doing is a little wild, then I don't know, you're probably just not taking enough, you're probably not swinging hard enough. Mixed like 18 metaphors in that answer, but you know, it's

如果你打击率 300，你就能进全明星赛；330，你就能进名人堂。

If you bat 300, you're going to the All-Star game, you know, 330, you're Hall of Fame.

是的，我的意思是，关于创造，我认为在棒球中，你最多只能打出全垒打，有一垒、二垒，你可以跑四个垒，对吧？但在创造中，上行空间几乎是无限的。你可以打出安打，也可以建造数十亿人使用的东西。它不是线性的。如果你最终建造了数亿或数十亿人喜欢的东西，那就能弥补很多不太成功的想法。但我觉得这没问题。这也是过程的一部分。我甚至不认为不成功的东西一定是失败，它们只是你在通往成功之路上的学习方式。我们都有很多学习方式。有一种市场化的方式，很多不同的初创公司做各种事，但在硬件领域，复杂的一点是尝试的成本很高，对吧？所以，我们想做的部分事情是，在我们开始制造这种眼镜之前，没有多少公司做这个。所以，很多实验基本上就是做很多不同的事情，看看什么有效，然后加倍投入，并尝试整合那些看起来有希望的东西，但可能我们做的某个产品中有一些有希望的元素，但其他方面整体感觉不太好。你把有希望的东西拿出来，整合到别的东西里，然后继续前进。所以，我不知道。我觉得这就是学习的方式。

Yeah, I mean, I think the thing though about creation is that in baseball, the maximum that you can do is get a home run and, you know, you have first base, second base, you can go four, right? In creation, the upside is almost infinite. It's not like you can do stuff, you can hit singles, or you can build something that billions of people use. And it's just not linear. You can like if you end up building something that hundreds of millions or billions of people like, then that ends up making up for a lot of ideas that didn't end up working that well. Now, but I think that that's fine. I like it's also this is all part of the process. I don't even think that the things that don't work aren't even necessarily failures as much as they are just ways that you were learning on the way to something that will work. You know, there's lots of ways that we all learn. You know, there's kind of a market-based way where you know, there's often a lot of different startups that do things, but in hardware, one of the things that's complicated is that it's kind of expensive to take shots, right? So, part of what we've wanted to do is, you know, before we got into building these kind of glasses, there weren't a lot of companies that were doing this. So, a lot of the experimentation we basically did by just doing a lot of different types of things and seeing what worked and doubling down on those things and trying to incorporate the things that seemed promising, but were but maybe there was like, you know, some product that we were doing, but there was something promising in it, but then a bunch of other things that made it not feel that good overall. You take the promising thing, you kind of incorporate into something else, and you keep going. So, I don't know. I think that that's just this is all like how you learn.

嗯，你谈到创造的规模是无限的。你知道，在过去的 20 年里，你做出的决策对人们的影响远超大多数 CEO。事实上，可能更接近于全球大国的领导人。这种压力让你付出了多少代价？

Well, you talk about sort of the scale of creation being infinite. And you know, in the position that you have been for the last 20 years, you make decisions that impact people in ways that are much bigger than most CEOs. And in fact, probably more akin to, you know, leaders of global powers. How much does that pressure cost you?

我认为不同的人有不同的思维方式。做出高风险但基于我信仰的决策，从来不会让我感到太大压力。让我感到压力的是，建造东西的乐趣之一在于与人合作。对吧？没有人是在真空中做这些的。我认为定义我生活中意义和职业目标的重要一点是，我想与最优秀的人合作，在新领域一起努力。这也是为什么设计眼镜如此有趣的原因之一，因为我并不是时尚人士，对吧？但现在我有很多东西要学。所以我们与优秀的公司合作，比如 EssilorLuxottica 是世界上最好的眼镜公司，我们从合作中学到了很多，我想他们也在技术方面从我们这里学到了东西。这是一段很棒的合作关系。与更多时尚界人士合作，以及在 Instagram 上做的工作，真的很有趣。我在个人生活中也尝试这样做，通过不同的爱好，比如为妻子设计珠宝之类的。设计不同的东西。我认为所有这些，与最优秀的人一起学习真的很有趣。对我来说，最压力大的事情是当与人合作不再有趣时。所以，如果你觉得没有良好的对齐，那才是真正让我压力大的事情。但对我哲学上相信的事情下大赌注，我不知道。也许这就是我的思维方式，但我对此没问题。

I think different people are wired differently. Making decisions that are high stakes, but are based on ideas that I believe in has never been something that has stressed me out that much. The thing that stresses me out, part of the joy of building things is doing it with people. Right? So, like no one's doing this in a vacuum. And I think a lot of how I define what is meaningful in my own life and what I want to do in my career is I want to work with people who are the best at what they do and try to work with them on in new areas. It's one of the reasons why, you know, I think designing the glasses has been so fun because I like I'm you know, I wasn't like a fashion person, right? And but now it's like but I have so much to learn, all right? It's like you know, so we work with, you know, the companies that are great and like EssilorLuxottica is the best glasses company in the world and we've learned so much working with them and I think that they've learned stuff from us on on the technology side. It's been awesome as a partnership and you know, working with more folks in fashion and the work that we've gotten to do on Instagram. It's really fun. I do this try I try to do it in my personal life with with different hobbies, some some of which, you know, we've talked about designing you know, designing jewelry for my wife, whatever. You know, it's like designing different things. I think all that stuff, it's really fun to just work with people who are the best at what they do to learn. The most stressful thing for me is when working with the people isn't fun. So, I think if you don't feel like you have good alignment, then that is actually the thing that really stresses me out. But taking big bets on things that I believe in philosophically, I don't know. Maybe that's just like how I'm wired, but like I'm fine with that.

我能接受自己的想法可能是错的，如果最终证明是错的，那确实会受伤。但我的天性就是几乎必须去做我相信的事。不去做那件事带来的痛苦，远大于为能否负担得起而焦虑或纠结。这就是我的性格。

I'm fine knowing that my idea could be wrong and it could hurt if it ends up being wrong, but I'm just wired in a way where I almost have to do what I believe. It will hurt me more to not do that thing than to stress or perseverate on whether I can afford to do it. That's just my constitution.

有没有谁的品味或直觉是你同等或更信任的？

Is there anyone whose taste or instinct you trust equal to or greater than your own?

有很多人我比自己的直觉更信任。在审美或关于美好未来的哲学上，我信任自己的直觉，但在我们工作的几乎任何领域，做那件事的人投入的时间都比我多。我同时在做很多不同的项目。无论是合作搭档还是内部团队，他们在那方面最好比我更有品味，否则我们到底在干什么？我自己做就行了。

There are a lot of people I trust more than my instinct. I trust my instinct on aesthetic or philosophy of what a good future would be, but in almost any area we work in, the people working on that thing spend way more time on it than I do. I'm working on a lot of different projects. Whether it's a partner I'm working with or the internal teams, they better have better taste on that than me, or else what are we even doing? I'll just do it myself.

当你进入一个想学习的领域时，你的自我教育过程是怎样的？

What is your process to educate yourself when you enter into a space that you want to learn about?

更像是沉浸式学习。不是那种问 20 个问题的正式流程。而是找一个我喜欢合作的人，或者几个人，因为大多数项目都很紧张且长期。和一群你不只是喜欢的人一起做——其实，我在创业初期得到过一个好建议。我 19 岁开始创业，以为应该雇有很多专业经验的人，但就是合不来。一位投资者说：“你应该雇那些你觉得想和他们做朋友的人。”这建议有点怪，但你和他们相处时间太长，真的需要彼此喜欢。所以关键是找到那些优秀且你想向他们学习的人。我现在有个原则：我只想雇那些在平行宇宙里我愿意为他们工作的人。这不是说我想换角色——我喜欢经营公司——但我认为我们团队的任何成员，我都能从他们身上学到很多，他们会做有趣的事。一旦找到对的人，你就会花很多时间在一起。这需要一定的谦逊，但复杂之处在于知道哪些是自己不知道的。这些合作不是单向的，而是双方都有所贡献。你要理解不同领域价值观的差异——比如设计眼镜和设计应用或手表。卓越在工艺、对细节的关注和大胆方面有相似之处，但也有细微差别。把握好这种平衡就是价值观，而当你喜欢并信任对方时，这种协商会更有效。

It's more immersion. It's not a formal process of asking 20 questions. It's finding someone I like working with, or multiple people, because most of these projects are intense and long-term. Doing it with a group of people you don't just like—actually, I got good advice early on building the company. I started at 19 and thought I should hire people with lots of professional experience, but it didn't click. One investor said, 'You should hire people you think you'd want to be friends with.' It's odd advice, but you spend so much time with them that you really need to like each other. So it's about identifying people who are excellent and you want to learn from. I have a rule now: I only want to hire someone if in an alternate universe I'd be okay working for them. That doesn't mean I want to flip roles—I like running the company—but I think any of our team members, I'd learn a lot from and they'd do interesting things. Once you find the right people, you spend a lot of time. There's some degree of humility, but the complexity is knowing which things you don't know. These collaborations aren't one-way; they're collaborations where each gives something. You try to understand how values differ across domains—like designing glasses versus an app versus a watch. Excellence has similarities in craft, attention to detail, and boldness, but subtle differences. Getting that balance right is values, and negotiating that works better when you like and trust the people.

除了享受和同事相处，你在这个位置上已经超过 20 年了。大多数同行要么退休，要么被淘汰。你在夏威夷有个漂亮的庄园，钱也花不完。是什么让你继续做这些，而不是干脆跑到岛上和家人一起消磨时间、做点小手工？

In addition to enjoying being around the people you work with, you've been in this seat for over 20 years. Most of your peers either retire or fall by the wayside. You have a lovely compound in Hawaii and infinite money. What keeps you doing this and not just absconding to the island to hang out with family and tinker?

普里西拉经常问我这个问题，尤其是事情不顺的时候。有意思的是——我们有个女儿有那种基因，就是有大量创意产出需要释放。说实话，在有孩子之前，普里西拉并不完全理解我这种天性。她以为我会干一阵子然后慢慢放松下来，但我不知道不干这个能做什么。那会很无聊。我可以休息几天，打打电脑游戏，但之后，就是有东西想放到这个世界里。我就是这样。花时间充电和放松很重要，和不同的人交流以获得多元视角，因为如果在一件事上钻得太深太久，创意能量会枯竭。但我不觉得自己会停下来。我会找不同的事情做。它们不一定都得是高风险的。在农场，有个项目是尝试生产世界上最高品质的牛肉——风险很低。我不卖它，但我对牛的基因很着迷，研究如何给它们提供最高密度的饲料。

Priscilla asks me that a lot, especially when things are hard. It's funny—one of our daughters has the gene where you just have a lot of creative output you need to make. Honestly, until we had kids, Priscilla didn't fully understand this element of how I'm wired. She thought I'd do stuff for a while and then mellow out, but I don't know what I'd do. That would be boring. I can rest for a couple of days, play computer games, but after that, there's just stuff you want to put into the world. That's how I'm wired. It's important to get time to recharge and relax, talk to different people for varied perspectives, because if you go too deep on one thing for too long, you can dry up creative energy. But I don't think I'm ever going to stop. I'll just find different things to do. They don't all have to be super high stakes. On the ranch, one project is trying to create the highest quality beef in the world—very low stakes. I'm not selling it, but I'm very into the genetics of the cattle, figuring out how to deliver the highest density diet to them.

我们开始种植夏威夷果，因为那种坚果密度极高。牛可以吃很多，所以它们会长胖更快，变得美味。但夏威夷果含油量高，需要烘烤。所以我们设计了一个烘烤坚果的流程来喂牛。然后我们希望它们吃得更多，结果发现酒精很有效，因为它能刺激食欲。这就是为什么高端牛肉会用啤酒喂养。但啤酒和水的比例多少合适？我不知道。让它们自己选。它们可以随意喝冰啤酒或常温水，自己调节平衡。所以现在我们一直在酿啤酒并供应。感觉停不下来了。

We started growing macadamia trees because that kind of nut is extremely dense. The cows can eat a lot, so they put on weight and become fat quicker and become delicious. But macadamia nuts have a lot of oil, so you need to roast them. So we had to design a process to roast the nuts to give them to the cows. Then we wanted them to eat more, and it turns out alcohol is great for that because it induces appetite. That's why very high-end beef is fed beer. But what's the right balance of beer versus water? I don't know. Let them choose. They get either as much cold beer as they want or as much room temperature water, and they can balance that. So now we're brewing all this beer and putting it out. It's like I'm never going to stop.

我喜欢。不管事情有多重要，我都会做一些不太重要的项目。

I like it. It doesn't matter how important or whatever the thing is. It's like yeah, I'll do some projects that don't matter that much.

你怎么找到时间的？三个孩子、妻子、九家公司，还有一个你在创新的养牛场。

How do you find time? Three kids, wife, nine companies, and a cattle farm that you're innovating on.

我不知道。我其实认为平衡的一部分是需要在不同的事情上找到平衡，因为如果你过于专注一件事，你会 burnout。也会让团队 burnout。这其实是我挣扎的一点，因为我容易着迷。但我不确定。我觉得就是和有趣的人一起做有趣的项目。这是美好的生活。

I don't know. I actually think part of the balance is you need to find some balance of different things to work on, because when you work on one thing too hard, you can burn out. And you can burn the people. So that's actually one of the things I struggle with, because I can get obsessive. But I don't know. I think it's just finding interesting projects to do with interesting people. It's a good life.

那么，Mark，非常感谢你的时间。这真的很棒。

Well, on that note Mark, thank you so much for your time. This has been really great.

嗯。

Mhm.